Identifying recurrent driver mutations in skin cancers by targeted UV damage sequencing - Abstract:
Skin cancers, such as melanoma, are among the most mutated human cancers due to the mutagenic action of
ultraviolet (UV) light. The high mutation load in skin cancers makes it challenging to identify driver mutations in
these tumors, particularly in non-coding DNA. For example, recent genome sequencing efforts have identified
dozens of recurrent exon mutations and hundreds of recurrent non-coding mutations in melanomas. However,
current evidence suggests that only a subset of these are under carcinogenic selection and contribute to
melanomagenesis. Many recurrent mutation sites are found in the DNA-binding sites of E26 transformation-
specific (ETS) family of transcription factors (TFs). We and others have recently shown that ETS and other TFs
induce high levels of UV damage, both in UV-exposed cells and in vitro, which can potentially explain the
presence of recurrent mutations at their binding sites. To test this hypothesis, we have developed new capture-
sequencing methods for mapping UV damage at targeted sites in the human genome. We propose to use
these methods to map DNA damage in UV-irradiated skin cells at sites of recurrent mutations in skin cancers.
The overall objective of this proposal is to use targeted UV damage sequencing to distinguish between
recurrent mutations that are simply caused by elevated levels of UV damage (often at TF binding sites) from
those that cannot be explained by UV damage, and therefore are more likely to be oncogenic mutations. In
Aim I, we will use the CPD-capture-seq method to map the formation and repair of UV-induced cyclobutane
pyrimidine dimers (CPDs) in UV-irradiated primary skin cells and cancer cell lines. In parallel, will develop a
new method, known as UVDE-capture-seq, to map less common 6-4 photoproducts (6-4PPs) and atypical
photoproducts at sites of recurrent mutations in UV irradiated cells. In Aim II, we will develop a bioinformatics
pipeline to use these capture sequencing data sets to identify candidate non-coding driver mutations by
triaging recurrent mutations associated with elevated UV damage. While it is possible that a few of the
recurrent mutations associated with elevated UV damage could contribute to carcinogenesis, multiple lines of
evidence suggest that the vast majority of these are passenger mutations. In parallel, we will functionally
characterize candidate non-coding driver mutations identified from our preliminary CPD-capture-seq data. We
will also determine whether UV damage induction due to transient ETS TF binding causes a subset of
recurrent exon mutations in melanoma and other skin cancers. Successful completion of these aims will
provide a new method for identifying driver mutations in skin cancers, particularly in non-coding DNA.
Importantly, the methods developed in this proposal could be adapted to analyze different classes of DNA
damage associated with recurrent mutation sites in other cancer types.